Staff – University of Copenhagen

IFSV English > Staff

Cox regression with missing covariate data using a modified partial likelihood method

Research output: Contribution to journalJournal article

Standard

Cox regression with missing covariate data using a modified partial likelihood method. / Martinussen, Torben; Holst, Klaus K.; Scheike, Thomas H.

In: Lifetime Data Analysis, Vol. 22, No. 4, 10.2016, p. 570–588.

Research output: Contribution to journalJournal article

Harvard

Martinussen, T, Holst, KK & Scheike, TH 2016, 'Cox regression with missing covariate data using a modified partial likelihood method' Lifetime Data Analysis, vol 22, no. 4, pp. 570–588. DOI: 10.1007/s10985-015-9351-y

APA

Martinussen, T., Holst, K. K., & Scheike, T. H. (2016). Cox regression with missing covariate data using a modified partial likelihood method. Lifetime Data Analysis, 22(4), 570–588. DOI: 10.1007/s10985-015-9351-y

Vancouver

Martinussen T, Holst KK, Scheike TH. Cox regression with missing covariate data using a modified partial likelihood method. Lifetime Data Analysis. 2016 Oct;22(4):570–588. Available from, DOI: 10.1007/s10985-015-9351-y

Author

Martinussen, Torben ; Holst, Klaus K. ; Scheike, Thomas H./ Cox regression with missing covariate data using a modified partial likelihood method. In: Lifetime Data Analysis. 2016 ; Vol. 22, No. 4. pp. 570–588

Bibtex

@article{13503998ecdb4d3c8e35c14ea0f932cd,
title = "Cox regression with missing covariate data using a modified partial likelihood method",
abstract = "Missing covariate values is a common problem in survival analysis. In this paper we propose a novel method for the Cox regression model that is close to maximum likelihood but avoids the use of the EM-algorithm. It exploits that the observed hazard function is multiplicative in the baseline hazard function with the idea being to profile out this function before carrying out the estimation of the parameter of interest. In this step one uses a Breslow type estimator to estimate the cumulative baseline hazard function. We focus on the situation where the observed covariates are categorical which allows us to calculate estimators without having to assume anything about the distribution of the covariates. We show that the proposed estimator is consistent and asymptotically normal, and derive a consistent estimator of the variance-covariance matrix that does not involve any choice of a perturbation parameter. Moderate sample size performance of the estimators is investigated via simulation and by application to a real data example.",
author = "Torben Martinussen and Holst, {Klaus K.} and Scheike, {Thomas H.}",
year = "2016",
month = "10",
doi = "10.1007/s10985-015-9351-y",
language = "English",
volume = "22",
pages = "570–588",
journal = "Lifetime Data Analysis",
issn = "1380-7870",
publisher = "Springer",
number = "4",

}

RIS

TY - JOUR

T1 - Cox regression with missing covariate data using a modified partial likelihood method

AU - Martinussen,Torben

AU - Holst,Klaus K.

AU - Scheike,Thomas H.

PY - 2016/10

Y1 - 2016/10

N2 - Missing covariate values is a common problem in survival analysis. In this paper we propose a novel method for the Cox regression model that is close to maximum likelihood but avoids the use of the EM-algorithm. It exploits that the observed hazard function is multiplicative in the baseline hazard function with the idea being to profile out this function before carrying out the estimation of the parameter of interest. In this step one uses a Breslow type estimator to estimate the cumulative baseline hazard function. We focus on the situation where the observed covariates are categorical which allows us to calculate estimators without having to assume anything about the distribution of the covariates. We show that the proposed estimator is consistent and asymptotically normal, and derive a consistent estimator of the variance-covariance matrix that does not involve any choice of a perturbation parameter. Moderate sample size performance of the estimators is investigated via simulation and by application to a real data example.

AB - Missing covariate values is a common problem in survival analysis. In this paper we propose a novel method for the Cox regression model that is close to maximum likelihood but avoids the use of the EM-algorithm. It exploits that the observed hazard function is multiplicative in the baseline hazard function with the idea being to profile out this function before carrying out the estimation of the parameter of interest. In this step one uses a Breslow type estimator to estimate the cumulative baseline hazard function. We focus on the situation where the observed covariates are categorical which allows us to calculate estimators without having to assume anything about the distribution of the covariates. We show that the proposed estimator is consistent and asymptotically normal, and derive a consistent estimator of the variance-covariance matrix that does not involve any choice of a perturbation parameter. Moderate sample size performance of the estimators is investigated via simulation and by application to a real data example.

U2 - 10.1007/s10985-015-9351-y

DO - 10.1007/s10985-015-9351-y

M3 - Journal article

VL - 22

SP - 570

EP - 588

JO - Lifetime Data Analysis

T2 - Lifetime Data Analysis

JF - Lifetime Data Analysis

SN - 1380-7870

IS - 4

ER -

ID: 160443327